> This doesn't seem to tally with my experience at all. Most shops will tell 
> you an item is out of stock and they typically have *no idea* when it will 
> be back in stock.

Most shops are no way near as organised as the likes of Tescos or any large 
manufacturing company.  And by no coincidence, they don't make as profit 
either...

> Or they promise it will be back in on date X, and it isn't, and then they 
> promise date Y, and it doesn't, and then they say a delivery is coming in 
> on date Z, but then they say it hasn't come in... You start to wonder how 
> they manage to do business at all!

Exactly.  But ask the store manager of Tescos when the next pallet of beans 
will be delivered and he will be able to tell you to the nearest 15 minutes.

> But anyway, while I can imagine a factory being able to predict how much 
> raw materials they need on a given day, predicting exactly how many tins 
> of baked beans you're going to sell today is obviously impossible.

Based on your sales of the previous days you can easily get an accurate 
prediction with confidence levels, eg 500 tins +/- 50 with 99% confidence. 
So you order 550 tins and you're 99% sure you'll be ok.  Sure, you'll have 
an average of 50 tins left over, but depending on that figure you can modify 
the quantity of the next order.  The optimising part is to keep the number 
of "left over" tins that you have to store to a minimum whilst ensuring the 
risk of running out of stock is close to zero.  You need to give a "cost" to 
the running out of stock scenario, and calculate what is the optimum.  Do 
not underestimate how much maths and stats is involved :-)

> Similarly, according to your scheme, if one delivery truck gets a flat 
> tire, or makes a wrong turn, or drives to the wrong store, or makes any 
> one of a million other simple mistakes, empty shelves result. Surely this 
> is no way to do business.

What is the chance of any of those things happening?  You estimate an actual 
number (eg by counting the number of flat tyres or times when the truck went 
to the wrong store in the past) and work out how much it is likely to cost 
you.  And then you see it is less than building and running a bigger 
warehouse (well I assume you do, as this is how most large companies work).

Post a reply to this message